AI Focus miss-focus sample

[Karl Gnter Wnsch]

I did some tesing on the film EOS-300 (V) with the same limitation as the EOS-300D will experience. I have especially looked at the issue of the AI Focus:

a) recomposing was working like a charm when following the same routine as I did on my trusty Fuji 602Z (One shot all of the time)

b) If I made the mistake (my behalf) to get significantly closer (getting the subject truly out of focus) to the focused subject and then recomposed the shot, there was a 50% chance of the camera going to AI Servo.

c) after some trying I could force the camera going to AI Servo when I wanted it to and stay out of it when I didn't.

My conclusion is:

a) although both systems might look similar maybe the Rebel/EOS-300 and the D10 might be using different sensitivities for the switchover.

b) there will be circumstances when I'd expect the system to fail. There is nothing stopping me switching to manual in time because I will be able to predict the failure most of the time. This is something to get used to but nothing that will get in the way too often.

c) there will be circumstances when this system will save my shot (I have plenty examples of out of focus shots of red deer that tell the tale of a lagging one shot AF)

Items b and c will probably even out the playing field for me.
--
regards
Karl Günter Wünsch

 

[MAC]

and I missed the test,

I know, I'm getting cryptic again Maybe a test like this:

Also some controlled gallery samples of Phil's wife driving a car
at 60 mph and Phil testing these cams with this test, how much
better does the science get??

--

Ulysses

--
MAC
http://www.digi-pictures.com

 

[Ulysses]

It is as Phil said earlier a little while ago: "It's like anything else, you have to understand it, allow for it and develop your technique."

I did some tesing on the film EOS-300 (V) with the same limitation
as the EOS-300D will experience. I have especially looked at the
issue of the AI Focus:

--

Ulysses

 

[Travis]

If the switch is fast enough (and Phil already indicated that there
is no real "delay" as such), then it doesn't matter a whole lot
whether it 'locks' into AI Servo, right?

Agreed, we really don't have anything concrete to show how much of a delay is introduced. On the other hand, if OOF shots can occur during the time it DOES switch, then I could see it being a problem if you take repetitive shots of action in a row.

Travis

 

[Travis]

I did some tesing on the film EOS-300 (V) with the same limitation
as the EOS-300D will experience. I have especially looked at the
issue of the AI Focus:

a) recomposing was working like a charm when following the same
routine as I did on my trusty Fuji 602Z (One shot all of the time)
b) If I made the mistake (my behalf) to get significantly closer
(getting the subject truly out of focus) to the focused subject and
then recomposed the shot, there was a 50% chance of the camera
going to AI Servo.
c) after some trying I could force the camera going to AI Servo
when I wanted it to and stay out of it when I didn't.

My conclusion is:
a) although both systems might look similar maybe the Rebel/EOS-300
and the D10 might be using different sensitivities for the
switchover.
b) there will be circumstances when I'd expect the system to fail.
There is nothing stopping me switching to manual in time because I
will be able to predict the failure most of the time. This is
something to get used to but nothing that will get in the way too
often.
c) there will be circumstances when this system will save my shot
(I have plenty examples of out of focus shots of red deer that tell
the tale of a lagging one shot AF)

Items b and c will probably even out the playing field for me.
--
regards
Karl Günter Wünsch

 

[Justme]

For those who don't know, BFS stood for Blue Flash Syndrome (you'll
have to do a search when it's working). I do know that Brian took
the issue to the fore-front, raising awareness of it.

Yes. Brian took credit for it. But the point being that the issue
was brought to the attention of Sony well before it was actually a
major conversation point in the forum.

My memories of those times are a bit vague right now. All I can recall is that some people thought he was trolling because he brought up this particular issue with the Sony F717. The thread was quite long from what I can recall and BFS was discussed in depth with Siverly trying to say it existed and showing examples and some people accusing him of trolling for doing so, etc. Then not too long after those sometimes heated debates started being posted in the Sony Talk forum, Sony offered a fix for it. All that matters I suppose is that it was corrected...for those who chose to take up the offer. I brought it up because you mentioned something which reminded me of the big BFS issue. Oh well, just a walk down ol' memory lane. We've been around too long

Quite a few
people in the Sony forum at the time, knocked him for it but they
ended up benefiting when it eventually caught the eye of Sony.

This is true. But Brian wasn't taken to task for that by some of
the forum members. What happened is that several folks allowed
their pride to get in the way of working together on it, including
Brian's. That was a loss for everyone. Still, the issue of BFS was
addressed and fixed for most that took advantage of it.

--

Ulysses

 

[MAC]

I use Custom Function 4,1 and then half shutter to meter and hold the * to focus in AISERVO, for still life, let off the * and recompose, for moving, kepp holding the , all in AISERVO.

you don't want to miss the bridal show.

These things get mentioned in the review.

And for those aspects and details that don't really need to have
an entire page devoted to them (like the 1.6x FOV crop), you have
the Learn center here at DPReview.com.

What folks need to learn is how to do their research. The info is
all there, and it's easy pickings. There's no need for unnecessary
redundancy.

--

Ulysses

--
MAC
http://www.digi-pictures.com

 

[MAC]

Brian was big reason for me skipping the 707. He uncovered real world issues that were not picked up prior.

MAC

For those who don't know, BFS stood for Blue Flash Syndrome (you'll
have to do a search when it's working). I do know that Brian took
the issue to the fore-front, raising awareness of it.

Yes. Brian took credit for it. But the point being that the issue
was brought to the attention of Sony well before it was actually a
major conversation point in the forum.

My memories of those times are a bit vague right now. All I can
recall is that some people thought he was trolling because he
brought up this particular issue with the Sony F717. The thread
was quite long from what I can recall and BFS was discussed in
depth with Siverly trying to say it existed and showing examples
and some people accusing him of trolling for doing so, etc. Then
not too long after those sometimes heated debates started being
posted in the Sony Talk forum, Sony offered a fix for it. All that
matters I suppose is that it was corrected...for those who chose to
take up the offer. I brought it up because you mentioned something
which reminded me of the big BFS issue. Oh well, just a walk down
ol' memory lane. We've been around too long

Quite a few
people in the Sony forum at the time, knocked him for it but they
ended up benefiting when it eventually caught the eye of Sony.

This is true. But Brian wasn't taken to task for that by some of
the forum members. What happened is that several folks allowed
their pride to get in the way of working together on it, including
Brian's. That was a loss for everyone. Still, the issue of BFS was
addressed and fixed for most that took advantage of it.

--

Ulysses

--
MAC
http://www.digi-pictures.com

 

[Sharif]

My conclusion is:
a) although both systems might look similar maybe the Rebel/EOS-300
and the D10 might be using different sensitivities for the
switchover.
b) there will be circumstances when I'd expect the system to fail.
There is nothing stopping me switching to manual in time because I
will be able to predict the failure most of the time. This is
something to get used to but nothing that will get in the way too
often.
c) there will be circumstances when this system will save my shot
(I have plenty examples of out of focus shots of red deer that tell
the tale of a lagging one shot AF)

Items b and c will probably even out the playing field for me.
--
regards
Karl Günter Wünsch

I second that opinion. nevertheless It annoyed me on my Rebel 2k that it sometimes switched into servo when I didn't want it to or did not switch when I wanted to. As you say it's not said that you switch manually when you have full control but you blame the camera when you can't switch yourself.

It happened to me only a couple of times during the 2 years I used the camera. Enough to be one of the main reasons for me to trade the Rebel into an Elan 7e. I'm happy that I did switch but don't ask me how many shots I missed because I forgot to switch into the right AF mode on the Elan. (Now I can't blame the equipment any more

alfalfa

--
http://members.chello.at/alfalfa/portfolio.htm

 

[Frank B]

I think when you hear the servo refocusing and see the refocusing in the viewfinder that it is pretty definitive that it has switched to servo mode. The camera continues to focus as you move it and when you hold it still the focus point does not light up or signal focus lock. If I had mistakenly let up the pressure on the shutter and it was still in Single Focus it would not focus as you moved the camera and when you repressed it would lock focus and signal.

Frank B

You need a good EXIF reader that exposes the Maker Notes.

I think this is the only definitive way to determine which method
the camera ACTUALLY chose.

I used a 28-135IS with IS shut off. I could here the servo start
to work and see the refocusing. I don't think the EXIF will show a
switch only that the focus mode was AI Focus.

--

Ulysses

 

[Frank B]

I would say in the majority of situations the camera will not refocus. However, I don't yet have a feel for the percentage.

Frank B

Do you have ANY situations where the camera decides NOT to refocus
that are similar to these?

I don't think I did that. I have been taking pictures using Single
Focus and recomposing for many years without this problem and don't
have it when I use Single focus with my 10D.

--
The Lowest Paid Concert Photographer Around
http://www.neonlightsimaging.com/artshow/final.htm
Photography -- just another word for compromise

'Since we can't keep crime in check, why don't we legalize it and
tax it out of business?' -- Will Rogers

 

[Ulysses]

My memories of those times are a bit vague right now. All I can
recall is that some people thought he was trolling because he
brought up this particular issue with the Sony F717. The thread
was quite long from what I can recall and BFS was discussed in
depth with Siverly trying to say it existed and showing examples
and some people accusing him of trolling for doing so, etc.

Oh man, those were classic times. But whenever you discover a real "bug" with a system, it always is. This thing about AI Focus reminds me somewhat of that, only it hasn't been determined to be buggy, simply not infallible. And there is no actual camera available off the shelf.

We've been around too long

One day we'll have to sit down and trade heart attack stories, too. )

--

Ulysses

 

[Ulysses]

How many times do I have to say....

Ohh, never mind.

The F707 is a distant memory now.

Brian was big reason for me skipping the 707. He uncovered real
world issues that were not picked up prior.

--

Ulysses

 

[Ulysses]

I need to think about that one for a bit.

Agreed, we really don't have anything concrete to show how much of
a delay is introduced. On the other hand, if OOF shots can occur
during the time it DOES switch, then I could see it being a problem
if you take repetitive shots of action in a row.

Travis

--

Ulysses

 

[Ulysses]

I guess what I'm saying is that we should verify it via the EXIF header. It ought to be recorded there.

What we hear may actually be part of the process as the camera is deciding what it wants to do (I'm not sure of this, and you'd honestly know better than I would as I never paid attention to it before).

But EXIF ought to verify. Thanks for your hard work and looking into all of this for our nutty questions, Frank.

I think when you hear the servo refocusing and see the refocusing
in the viewfinder that it is pretty definitive that it has switched
to servo mode. The camera continues to focus as you move it and
when you hold it still the focus point does not light up or signal
focus lock. If I had mistakenly let up the pressure on the
shutter and it was still in Single Focus it would not focus as you
moved the camera and when you repressed it would lock focus and
signal.

--

Ulysses

 

[Frank B]

I will try it and see if I can detect a difference. I did use Tv mode for my "test" pictures.

Frank B

If you assume that my 10D is not defective and that I did not mess
up the test and that AI Focus does sometimes switch to servo when
recomposing then the question is how often will it make a
difference. Maybe not too often.

1. I suspect it only switches to servo when the center focus point
after re-focus is either closer or further away and it does that
only occasionally.

2. If it does switch to servo and the two focus points are within
depth of field the subject will still be sharp, although not quite
as sharp as it would be if the camera had focuses accurately.

I need to test further whether when it switches to servo mode it
has trouble switching back to Single Focus when you half press the
shutter for the next picture (as reported in the Luminous Landscape
review). If it does not switch back to Single Focus that is
another a problem. I think my camera switches back without a
problem, but I am not sure.

Frank B

 

[Ulysses]

Sounds kinda fiddly.

Is it, in the real world?

I use Custom Function 4,1 and then half shutter to meter and hold
the * to focus in AISERVO, for still life, let off the * and
recompose, for moving, kepp holding the , all in AISERVO.

you don't want to miss the bridal show.

--

Ulysses

 

[Sharif]

Sounds kinda fiddly.

Is it, in the real world?

it sounds more complicated than it is. Actually many photographers use this method with their Canon EOS SLR's.

It's simply the separation of the AF from the shutter button.

regards,
alfalfa

I use Custom Function 4,1 and then half shutter to meter and hold
the * to focus in AISERVO, for still life, let off the * and
recompose, for moving, kepp holding the , all in AISERVO.

you don't want to miss the bridal show.

--

Ulysses

--
http://members.chello.at/alfalfa/portfolio.htm

 

[DavidP #28649]

That's how I use my 1D now, almost all of the time.

It takes a few times out with it to get used to it. But it's become second-nature to me now, and I don't know how I lived without using it before.

Sounds kinda fiddly.

Is it, in the real world?

--
The Lowest Paid Concert Photographer Around
http://www.neonlightsimaging.com/artshow/final.htm
Photography -- just another word for compromise

'Since we can't keep crime in check, why don't we legalize it and tax it out of business?' -- Will Rogers

 

[MAC]

to do this finger, thumb... yoga I need to get Big Ed and try it that way.

exposure is key and metering is key.

half shutter, meter on face or neutral grey, and recompose

Then use your thumb to hit the * to focus using center point that is the most reliable...then remove the * to lock focus and recompose....shoot.

Or for moving subject keep the * engaged and shoot.

This is technique used by DavidP, pekka and me, allows you to be in AISERVO all the time and be ready to

Blast away, blast away, blast away into the night

MAC

Is it, in the real world?

I use Custom Function 4,1 and then half shutter to meter and hold
the * to focus in AISERVO, for still life, let off the * and
recompose, for moving, kepp holding the , all in AISERVO.

you don't want to miss the bridal show.

--

Ulysses

--
MAC
http://www.digi-pictures.com